Assessment of two essential elements of BVDV control on selected Flemish dairy and beef farms

Bovine viral diarrhea (BVD) is an infectious disease of cattle with a worldwide distribution (Ridpath, 2010a), causing significant economic losses (Houe, 1999; Fourichon et al., 2005). Infections can be either persistent or transient. Persistently infected (PI) cattle are key in spreading and maintaining the infection within and between herds, as they continuously shed large amounts of BVD virus (BVDV) during their entire lives (Lindberg and Houe, 2005; Fulton et al., 2009). Therefore, most transient infections are caused by direct contact with PI cattle. The direct consequences of transient BVDV infection may vary from subclinical or mild disease to acute outbreaks with severe disease and high mortality. Moreover, transiently infected (TI) cattle may suffer from immunosuppression, which makes them more susceptible to secondary infections (Brackenbury et al., 2003; Ridpath, 2010b; Chase, 2013). Between Assessment of two essential elements of BVDV control on selected Flemish dairy and beef farms


INTRODUCTION
Bovine viral diarrhea (BVD) is an infectious disease of cattle with a worldwide distribution (Ridpath, 2010a), causing significant economic losses (Houe, 1999;Fourichon et al., 2005).Infections can be either persistent or transient.Persistently infected (PI) cattle are key in spreading and maintaining the infection within and between herds, as they continuously shed large amounts of BVD virus (BVDV) during their entire lives (Lindberg and Houe, 2005;Fulton et al., 2009).Therefore, most transient infections are caused by direct contact with PI cattle.The direct consequences of transient BVDV infection may vary from subclinical or mild disease to acute outbreaks with severe disease and high mortality.Moreover, transiently infected (TI) cattle may suffer from immunosuppression, which makes them more susceptible to secondary infections (Brackenbury et al., 2003;Ridpath, 2010b;Chase, 2013).Between herds and animals, substantial differences in clinical presentation of BVDV infection have been noticed.This variability has been attributed to herds/animals having different immune statuses (Lindberg, 2003), or differences in strain virulence (Bolin and Ridpath, 1992;Pellerin et al, 1994;Baule et al., 2001;Walz et al., 2001;Kelling et al., 2002).Because of the marked variation in clinical presentation, it is usually difficult to detect the presence of a BVDV infection in a herd by its clinical presentation alone (Lindberg and Alenius, 1999;Ridpath, 2003;Evermann and Barrington, 2005).As a result, monitoring by diagnostic testing for the presence of BVDV is vital to determine the herd status and for effective BVDV control.
In the absence of a nationwide eradication program in Belgium, control is typically performed at the herd level by decision of the farmer.Meticulous tracing, correct administration (Laureyns et al., 2010) and the implementation of the key principles of BVDV control as stated by Lindberg and Houe (2005) are essential to successfully control BVDV at the herd level.These principles are: stringent biosecurity, the detection and removal of PI cattle, continuous monitoring and the potential implementation of vaccination.As the prevalence of BVDV infection at the farm level is high in Belgium (Sarrazin et al., 2013), biosecurity measures are highly important to protect herds against BVDV (re)infection.Therefore, the implementation of biosecurity measures has to be the first step in BVDV control in Belgium.Most importantly, PI animals should be removed from the herd, as they play the key role in the transmission of BVDV by continuous shedding of infectious virus (Lindberg and Houe, 2005).In BVDV-free herds, longitudinal surveillance should be combined with biosecurity enhancements to detect and prevent potential (re)infection, and allow prompt action in the event of disease incursion.Monitoring can be performed using serological spot tests (Houe, 1994;Booth and Brownlie, 2012), and has to be continued as long as BVDV is present in the region.Although not 100% effective in protecting every individual animal, vaccination can be an essential component of a herd level control program (Ridpath, 2012).However, if eradication is to be achieved, vaccination must be combined with the other three principles of BVDV control (Lindberg and Houe, 2005;Rodning et al., 2010;Booth and Brownlie, 2012).
Although it has been emphasized in many publications that BVDV control requires a combination of different, strictly executed measures, little information is available on how these recommendations are implemented in the field.This study highlights the BVDV management on Flemish farms.

MATERIAL AND METHODS
In 2011, a large multicenter study was conducted (Pfeiffer et al., 2012) to identify calf-level factors associated with bovine neonatal pancytopenia (BNP) on BNP-affected farms (Jones et al., 2013), and herd-level factors that explained why some farms experienced cases of BNP and others did not.Since BNP has been hypothesized to be associated with BVDV vaccination, a substantial part of the interview was on BVDV management.Questions of relevance for the present study were selected from the BNP questionnaire and are presented in Table 1.
The Belgian contribution took place in the Flemish speaking part of the country, Flanders, and was conducted by the Faculty of Veterinary Medicine of Ghent University in cooperation with the Flemish Animal Health Organization "Dierengezondheidszorg Vlaanderen" (DGZ).Through a call on the website of DGZ and notifications in veterinary and farmer magazines, veterinarians and farmers were encouraged to report suspected BNP cases.Case farms were visited for sample collection, and a questionnaire was used to interview the farmers on colostrum feeding, cattle health management, disease management and medication use.Farms where the veterinarian had reported a suspicion of BNP, were classified as case farms.The latter farms belonged to the clientele of the same veterinaty practice, and had never been diagnosed with BNP before.The control farms were of the same type and approximately the same size as the case farms.Farmers of the control farms were interviewed by telephone, using the same questionnaire.All of the herd managers of the case and control farms were interviewed by the first author.Data of 241 Flemish farms was available for the study.
The recognition of BVDV infection by its clinical presentation alone is very difficult, if not impossible.Therefore, in this study a herd was only considered BVDV-free if the status was based on a test-and cull-program or a herd test.A test-and-cull program consists of virological testing by PCR or antigen-ELISA of all cattle in the herd followed by the culling of PI animals.In the present study, the continuous testing of all newborns by antigen-ELISA for already more than one year, or regular serological spot tests of which the last one took place within the last twelve months, were both accepted as a herd test.The intention of using a spot test is to detect BVDV circulation in a herd by testing five to ten blood samples of young stock between eight and twelve months old for the presence of BVDV antibodies (Houe, 1992).

RESULTS
The results are summarized in Table 2.The farmers were asked 'Was the herd BVDV-free for the last 12 months?'.Of the 241 herds, 82 had a known BVDV status that was based on testing; 66 of those 82 had been BVDV-free and 16 had been BVDV-infected during the past year.On 158 of the farms (66%), there had been no monitoring for BVDV.When the farmers of the herds that were BVDV-free, were asked for how long they had held this status, eleven of the herd managers communicated a date within the last twelve months, although they scored their herd BVDV-free for the whole of the preceding twelve months in another answer.
On 55 (67%) of the 82 herds with known BVDV status (infected and free farms), BVDV testing consisted of a herd test while on the other 27 farms, a test-and-cull method was in use.
On the question 'Did you have a confirmed PI animal on your farm within the past twelve months?',twenty of the 241 herd managers answered positively (8%), whereas the majority (153) did not know whether a PI animal had been on the farm or not, as they declared that a monitoring program for BVDV had not been in place (63.5%).On 68 (28%) of the herds, the BVDV status was known, but no PI animals had been detected within the last twelve months.
On was still continued at the time of the interview (farms that only vaccinated young stock were excluded); 66 of these 93 (71%) herd managers did not know the BVDV status of their herds.
Of the twenty herds where a PI animal had been found within the previous twelve months, seven farmers applied the BVDV vaccination of adult cattle at the time of the interview.Of these seven herds, five had been vaccinated for two years or longer.

DISCUSSION
The main objective of this study was to investigate the BVDV management on selected farms by describing common policies for PI animal detection and monitoring, two of the three essential BVDV control measures.The results were collected as part of a larger case-control study on the identification of risk factors for the occurrence of BNP, a BVDV vaccination related disease (Jones et al., 2013).
The fact that all of the questions for this study were asked by the same person, both on case and control farms, reduced the likelihood of bias inclusion.Although the same questionnaire was used for both case and control farms, on the case farms, the interviews occurred face to face, whereas on the control farms, they were conducted by telephone.Therefore, the answers collected on the case farms might be considered more reliable, as they were better supported by written or electronically stored data such as lab results provided by the farmer at the herd visit.Two different descriptions of the same question indicated that the BVDV status was unknown on 66% (n=158) and 63.5% (n=153) of the farms during the past year, respectively referring to the absence of BVDV and the presence of a PI animal.The fact that there was little difference between the figures obtained via both questions reinforces the certainty that the majority of the herd managers did not know the BVDV status of the herd.It is interesting to note that when asked if they knew their BVDV status, 16 farm managers answered that their herds had been infected in the last 12 months.Yet, when asked about the presence of PI animals, twenty farm managers stated that PIs had been identified on their premises in the last twelve months.Remarkably, four farmers did not appear to know that the presence of PI animals is an indicator of herd-level infection.Moreover, when asked for how long their herds had been BVD-free, eleven herd managers communicated a date within the last twelve months, although they had previously scored their herds BVDV-free for the whole of the preceding twelve months.These examples demonstrate that a question may produce different answers when asked in a different way, and illustrate potential difficulties when working with questionnaires.
Interestingly, two veterinary practices of the 43 involved were responsible for 16 of the 55 herds with BVDV surveillance.This result suggests that few veterinary practices implement BVDV monitoring in their herd health management programs.
Farms were not classified as BVDV-free if this status was obtained from clinical signs.The only methods accepted for defining the BVDV status were: a herd test or a test-and-cull program during the last year.Of the 82 herds with known BVDV status, those 27 where a test-and-cull program had been used, may not necessarily be considered as herds with BVDV surveillance.On these farms, the herd managers had performed one whole herd test for the presence of PI animals during the last year, reacting on a suspicion of BVDV infection, and afterwards assumed that the herd had been BVDV-free since then.The only methods that may be considered as surveillance are regular serological spot tests or virological testing of all newborn calves on a whole blood sample from the age of two months onwards.Ear notch testing (Kuhne et al., 2005;Hill et al., 2007) may be an alternative, but was not yet available in Flanders at the time the study was conducted.It can be concluded that only 23% (n=55) of all the herds implemented a BVDV surveillance program in the strictest sense.Nevertheless, virological testing of all newborns as a sole monitoring measure might not rapidly identify re-infection of the herd (Houe et al., 2006), for instance because of overlooking a PI calf due to an administrative failure (Laureyns et al., 2010), a false negative result (Presi et al., 2011;Fux and Wolf, 2012) in the previous detection and removal of PI cattle or external re-infection.This is especially true if testing all newborns has only been conducted for a short period of time, and does not follow a testand-cull program, since PI animals can effectively remain within a herd until the point at which they calve.
The BNP case farms were all BVDV-vaccinated, and the corresponding controls were linked to the same veterinary practice.Therefore, it can be supposed that the veterinarians involved were conscious of the importance of BVDV infection, because they had advised vaccination to their clients.On the other hand, it appeared that most of them did not apply all principles of BVDV control (Lindberg and Houe, 2005), as on 71% of those farms still vaccinating at the moment of the interview, the BVDV status was unknown.This raises the question as to whether the veterinary profession as a whole still has an overreliance on BVDV vaccination for control of the disease, when it should be considered as only one component of a range of control measures that should be implemented.
As the multi-country study was focused on BNP, a BVDV vaccination-related disease, the number of vaccinating herds may have been overestimated.Most likely, the control group had been vaccinated as well, since it was recruited from the same veterinary practice as the case herds, and consequently, the same BVDV control strategy may have been implemented.On the other hand, at the time of the herd visit, most of the farmers had already been informed on BNP.Some had changed to another BVDV vaccine, whereas twenty of the original 155 vaccinating herd managers had ceased vaccination at the time of the interview, most likely for fear of BNP.As a result, the figures are not suited to interpret the BVDV vaccination prevalence in Flanders.They rather show some shortcomings of vaccination management.Although vaccination on its own is not sufficient to eradicate BVDV from a herd (O' Rourke, 2002;Booth and Brownlie, 2012;Ridpath, 2013), 66 of the 93 herd managers who were still vaccinating against BVDV at the time of the interview, did not know their BVDV status (71%).Vaccination should be combined with all three other necessary components of BVDV control: biosecurity measures, the detection and removal of PI animals, and monitoring (Lindberg and Houe, 2005).Moreover, it should be realized that the use and application of BVDV vaccines in the field are not always correct (Meadows, 2010).The present study shows that in almost one third of all the vaccinating herds (27%; n=42), only young stock under six months of age had been vaccinated, while the advice is to reach a 100% coverage of the adult herd (Lindberg and Houe, 2005) with the main objective of preventing infection of pregnant cattle.
On 83 of the 155 vaccinating herds (53.5%), the decision of starting BVDV vaccination was made at the occurrence of a BVDV problem.Another 10% (n=16) of the herd managers did not really have the intention to control BVDV.They used a trivalent vaccine containing a BVD-component, to protect calves from bovine respiratory disease (Rispoval®3-BRSV-Pi3-BVD Pfizer Animal Health), but did not have a BVDV vaccination program for older cattle.Not surprisingly, among these 16, there was only one dairy herd and 15 beef farms, since Belgian Blue cattle, the predominant beef breed in Belgium, are substantially more susceptible to respiratory disease than dairy breeds (Bureau et al., 1999).Nevertheless, 42 of the farmers vaccinated only three-month-old calves.Since only 16 farmers stated that they vaccinated against respiratory disease, it can hence be assumed that 26 other herd managers considered vaccinating only three-month-old calves as a herd level BVDV control strategy.
Only 35.5% (n=55) of the vaccinating farmers stated to have started the vaccination program to protect the herd from BVDV infection without previous BVDV problems.Obviously, both these farmers and the ones who started vaccination after their herds had suffered from BVD problems (53.5%),i.e. the majority of vaccinating farmers, must have been aware of the economic consequences of BVDV infection.
Finally, the observation that a PI animal had been detected on five farms, despite BVDV vaccination during the last two years or longer, shows that continuous monitoring is necessary, even for vaccinating herds.

CONCLUSION
This study illustrates that even on selected farms, where many farmers were willing to vaccinate against BVDV and thus conscious of the impact of BVDV infection, the necessary elements of BVDV control were insufficiently implemented.In particular, too many control strategies were based on vaccination alone and only few herds were monitored for BVDV.These findings suggest that on many Flemish farms, BVDV control remains incomplete and consequently inefficient, despite repeated communication and education from regional animal health services and veterinary faculties.It is up to the veterinarians to train their clients to control BVDV efficiently.

Table 1 . Questions of the bovine neonatal pancytopenia questionnaire selected for the present study.
155 (64.5%) of the 241 farms, a BVDV vacci-appeared to be vaccinated, all with a trivalent vaccine containing a BVDV component (Rispoval®3-BRSV-Pi3-BVD, Pfizer Animal Health).Of these 42 herd managers, only 16 declared that the reason for applying Rispoval®3-BRSV-Pi3-BVD vaccination had been purely for the prevention of respiratory disease.On 93 farms, the vaccination of heifers and/or adults